Effect of Surface Pavement Type on Traffic Noise

T his paper investigates the issue of surface-type effects on traffic noise in Baghdad. Since the raw materials for both flexible and rigid paving are available from local sources, the decision on selecting the type of paving which depends on the budget of the project and the road's importance and function. Knowing that for high traffic volumes and a high percentage of heavy vehicles, rigid pavement is more suitable compared to flexible pavement. In Baghdad, some highways consist of flexible pavement and others of combined pavement (flexible segments and rigid segments), so the study of the effect of surface type on traffic noise becomes an important matter. This study selected three highways: one with flexible pavement and two with the combined pavement. The results showed that the traffic noise generated by traffic on rigid pavement generates more noise than traffic on flexible pavement by about 2 dBA. A prediction model was generated to predict traffic noise depending on five variables (vehicle speed, traffic volume, skid number, distance between the edge of the inner lane and the sound level meter, and surface type). The measured traffic noise levels ranged from 73.23 to 82.86 dBA, which exceeded the permitted limits compared to the permissible Iraqi standards.


INTRODUCTION
Road traffic noise is a complicated phenomenon due to the involvement of several variables.The traffic noise levels generated depend on the traffic flow, the speed of the stream, and the heavy vehicle percentage (Ali and Albayati, 2022).The source of generated noise from an individual vehicle is divided into two sources, first, from the vehicle itself, such as engine noise, exhaust systems, and engine cooling systems; while the second source is from the air dynamics of the vehicle while moving and the interaction of tire-pavement (Ali et al., 2018).Action is taken to reduce the traffic noise from its sources, such as using hybrid vehicles and electrical engines (Li et al., 2016).The road pavement surface was also studied, and several techniques were invented that showed adequacy in decreasing traffic noise, such as porous asphalt pavement for flexible pavement roads and grinded surfaces for concrete pavement roads (Skarabis and Stöckert, 2015).Other parameters affect the traffic noise levels, such as the distance between sound receivers and the road, the geometry of the roadside, the ground cover for the area beside the roadway, the availability and type of the noise barrier, and the weather conditions (Azodo et al., 2019).Economic development, industrialization, and urbanization increased not only highway traffic but also elevated noise levels.As a result, it is now required to assess the features of traffic noise in relation to different types of pavement surfaces, vehicle types, and vehicle speeds (Cho and Mun, 2016).
Determining the traffic noise levels that may occur or have already occurred is crucial during the planning and designing stages and after constructing new roadways.Thus, noise prediction and mathematical dispersion models are applied.The method used to measure noise levels varies by country.Many more developed nations have created their own models for the prediction and dispersion of noise.Models from other countries are used by countries without their own models, whether or not they have been adjusted for local conditions (Pozder, 2012).Parnell & Samuels, 2006 studied the influence of various types of road pavement surfaces on traffic noise in Australia.The results of the study showed that the quietest pavement was stone matrix asphalt pavement, dense graded asphalt pavement, chip-sealed pavement, and concrete pavement, respectively (Samuels and Parnell, 2007).
Seung & Mun, 2008 studied the effect of various types of road pavement surface along the southbound side of the Jungbu Inland Expressway in South Korea.The location studied was divided into nine segments, and the vehicles were classified into eleven classes.The analysis of measured data showed that the asphalt pavement surface has higher acoustical benefits than the concrete pavement surface for passenger and heavy vehicles (Cho and Mun, 2008).Lakušić and his coworkers investigated the effect of different surface pavement types on road traffic noise at the Llova Bridge in Croatia (Lakušić et al., 2010).The former road surface type was concrete, rehabilitated by placing an asphalt surface.Various variables were considered during the data analysis, such as vehicle speed, surface condition, and surface evenness.The results showed that the noise levels measured on the concrete surface were higher than the traffic noise on the asphalt concrete surface due to the evenness of the surface and the incompetent execution of the expansion joints of the concrete pavement.In Bosnia and Herzegovina, Pozder, 2012, studied the effects of the pavement surface on traffic noise levels.The data measurement was at 70 sites on the road network.The result showed the generation of a traffic noise prediction model considering the surface condition (Pozder, 2012).Awwal and others studied the traffic noise levels on the Skudai-Pontian Highway in Johor, Malaysia (Awwal et al., 2021).The study aimed to assess the noise levels on two types of pavement.The traffic characteristics were measured, as well as the traffic noise.The data were used to calculate the predicted noise levels using five different prediction models.It was found that all models did not suit the observed noise, except the Penang noise model.The model was developed to fit the predicted traffic noise for specific conditions.Therefore, it can be deduced that pavement type and traffic condition impact traffic noise, and creating a customized traffic noise model for each pavement and traffic condition can assist in planning future changes in traffic noise circumstances.The current study is interested in assessing roadside noise levels on existing roadways with asphalt and concrete pavements.Also, it includes a prediction for the noise level in terms of vehicular stream characteristics, i.e., speed, volume and distance from the noise source besides the surface texture of pavement as evaluated by the skid number.The prediction model has been developed using MINITAB V16 software.

ROAD SURFACE PAVEMENT
Road surface pavement can be divided into two types: asphalt concrete pavement and concrete pavement.

Asphalt cement pavement
This type of pavement typically consists of a base course, binder course, and surface course layers.Fig. 1 shows a typical asphalt-cement cross-section of a road.Hot mix asphalt (HMA) is the typical type of surface layer.The layer is produced by mixing hot asphalt with graded aggregate and mineral fillers such as limestone and Portland cement (Albayati, 2018).Many modifications occurred, and the results were warm mix asphalt, porous asphalt, and opengraded hot mix asphalt (Bendtsen and Andersen, 2005).(1) where:  50 is the sound level exceeded for 50% of the time of the measurement period (t), dBA.
is the hourly volume (veh/hr).
is the distance from the center of the inner lane to the observation point in feet.
Nickson, 1965 model's condition was that the mean speed of vehicles is 40 mph and 10% heavy vehicle percentage.Lamure, 1965 made developments on Nickson's model to be applied for traffic volumes of (1200-5000) veh/hr and a heavy vehicle percentage of 15%.
50 = 50 + 10 (/) (2) In England, Johnson and Saunders, )1968), cited in (Kamil and Al-jameel, 2022), developed a prediction model with more complexity.The limitation of the model is that the heavy vehicle percentage is 20%.They also studied the effect of ground cover and the road gradient, and correction factors were considered.Eq. ( 4).shows the Johnson and Saunders model.
50 = 3.5 + 10 ( 3 /) In 1972, Wesler developed a method named TSC.It provided a computerized traffic noise prediction method.The concept of the model was to calculate the noise generated by light vehicles and heavy vehicles separately.Then the summation represents the total noise.Eq. ( 5).represents the TSC model.while for roads and highways in urban areas when the hourly volume is less than 1000 veh/hr the following equation is used: In Italy, a prediction model was developed by the "Consiglio Nazionale Delle Ricerche" and then improved by Cocchi.The model is based on a linear source relation between traffic parameters and noise levels.Traffic noise can be calculated from Eq. ( 10) (Quartieri et al., 2009).
=  + 10 (  +   ) − 10 (/   ) +  (10) where:   is the hourly volume for light vehicles (veh/hr)   is the hourly volume for heavy vehicles (veh/hr)  0 is the reference distance of 25 meters (m) is the distance between the lane center and the observation point on the road's edge (m).is the adjustment factor , is parameters affected by the road and vehicle type (in Italy, α=35.1 and β=6)

Location Selection
The locations selected in the study tended to cover all concrete pavement highways and Baghdad city's main asphalt pavement highways.The studied locations are given in Table 1.Fig. 3 shows the map location for the selected roads.

Devices Used in the Study
During the process of measuring the road traffic noise, the following devices were used: 1. Sound level meter, using a CEM (DT-8852 DATA LOGGER) model device.
2. Speed Gun detector, using a Bushnell model velocity speed gun for detecting vehicles' speed at locations.3. British pendulum tester for skin number measurements.

Test Procedure
In each location, the sound level meter was set at the height of 1.5 meters.The vertical distance between the edge of the inner lane and the sound level meter was 1, 3, 6, and 9 meters.The sound level recording duration was one hour.Vehicle spot speed was measured with a speed gun, pointed in the vehicle's direction and manually recorded simultaneously with the sound levels.The speed and frequency were analyzed, and the 85-percentile speed was computed and used for each study hour.Traffic volume was also counted manually during the same duration, taking into consideration the vehicle type.The vehicles were classified into two categories: passenger cars and heavy vehicles.Since all locations with flat terrain, a passenger car equivalency factor of 1.5 was used to convert the heavy vehicles into equivalent passenger cars.Every location was tested in the three-time interval, 5-6 AM, 7-8 AM, and 9-10 PM, during weekdays only.The weather conditions during the data collection were clear, with low wind speeds.In each location, the selected data collection section avoided any traffic streams obstructions such as intersections, interchanges, bumps, pedestrian crossings, and surface distortions and had no acceleration or deceleration lanes.Fig. 4 shows the measuring procedure.

RESULTS AND DISCUSSION
The road traffic noise was measured in the selected locations, as shown in Table 2, where average values for several measurements for PCC and HMA road surface type locations were taken.The average values of Leq for HMA and PCC roads are given in Table 3.A comparison between the measured Leq(average) for different types of surfaces and the Iraqi standards is conducted as given in Tables 3 and 4.
where Leq(average) is calculated as: where N is the number of readings It is found that the noise levels measured from PCC surfaces exceeded the that measured from HMA surfaces.The measured noise levels in all locations were higher than the permissible levels limited by Iraqi standards.The results are shown in Fig. 5.A prediction model is generated per the measured data (sample of them) given in Table 5.The model includes the vehicle speed, traffic volume, skid number, distance between the inner edge of the road and the sound level meter, and the type of road surface.The data were analyzed using the Minitab 16 software program, and the resulting model is given as: = 61.8+ 0.215 + 0.00206 − 0.773 − 0.127 − 0.847 where:  is the 85-percentile speed for the stream (km/hr.). is traffic volume (pc/hr).
is distance between noise level meter and inner edge of the road (m).
is skid number for road surface. is the type of road surface (assumed 1 for PCC and 2 for HMA).

CONCLUSIONS
This paper studied the traffic noise on three different highways in Baghdad city.Two of the selected highways consist of two segments with different surface types (PCC and HMA).The results showed that the traffic noise generated was higher for PCC roads as compared to the traffic noise for HMA roads.Traffic noise levels measured ranged from 73.23 to 82.86 dBA.
The measured noise levels showed to be higher than the limits permitted by the Iraqi standards for all locations.For PCC highways, the noise levels were 32.65 % and 58.54 % higher than the permissible levels for daytime and nighttime, respectively.For HMA highways, the noise levels were 28.40% and 55.64% higher than the permissible levels for daytime and nighttime, respectively.A prediction model was generated depending on five variables: vehicle speed, traffic volume, skid number, the distance between the sound level meter and the inner edge of the road, and surface type.The generated model shows an acceptable ability for prediction since R 2 was found to be 86.26%.

Figure 3 .
Figure 3. Map for the selected location in Baghdad

Figure 4 .
Figure 4. Data measurement procedure for concrete and asphalt pavement

Figure 5 .
Figure 5. Leq (Average) for different road surface types

Table 1 .
Locations selected for the study

Table 2 .
Average Leq measured for selected locations

Table 3 .
The average   for each type of pavement

Table 4 .
The Iraqi specification limits for sound levels (

Table 5 .
A sample of measured data for location 5, SN. of 54